Guru
Joined: 13/03/2017
Location: AustraliaPosts: 460
| Posted: 06:47am 02 Mar 2019 |
Poida
Great post mate. Amazing timing - I've been working on dvdt this week. What I have found is that dvdt can cause minor shoot-through at quite slow changes in Vdrain (and less so Vsource). Much slower that the data sheets suggest. I have included in my driver board a clamping FET which directly clamps the Gate (200 mOhm) to ground and it will still experience some turn-on, via Cgd. I can reduce the shoot-through, say of the lower fet, by slowing down the turn on action of the upper fet, particularly until Vthreshold. I did this by adding an inductor in series with Rg, which I also increased in value. You could also do this by adding more inductance to the load.
I am very jealous of your CRO, and diff probes. Which diff probe are you using?.
I am about to try a different drive circuit where I clamp the gate to -5 or -12 volts. The reason I'm doing this is that it will take a lot more charge from Cgd before Cgs charges up to its threshold voltage.
I believe it is important to get this right at no load, because High loads are likely to be inductive and will slow dvdt, its at no load that dvdt will show it affects and possibly will cause destruction of the mosfets.
Cheers
Andrew